Backed by Y CombinatorFour documents and two sheets that catch when a single time zone hides a negative arrival gap, and when a manager's answer stops matching everyone else's.
Free download · No account needed
What you get
Pack of 6 files: 4 documents, 2 spreadsheets
Internal Announcement Template for ChangeAudience Sequence
Every office and every distributed asset, checked against its own next working hour, not headquarters'.
| Audience or Asset | Local Time | Governing Deadline | Gap (Hours) |
|---|---|---|---|
A positive send gap and a negative arrival gap can be the exact same broadcast. This sheet computes both, per office, before anything goes out.
Thorncliff Robotics's first draft sequenced its layoff announcement against Austin's clock alone: talking points, then affected-employee 1:1s, then the all-hands, each thirty minutes to an hour ahead of the external release. Every send gap came back positive. Converting the same three sends into local time for the company's other three offices found three of them comfortably inside a normal waking day, and one that was not. The all-hands landed at midnight in Sydney, where the external release was already public six hours before anyone there had a realistic chance to see it.
The exit plan also triggered a Form 8-K under Item 2.05, Costs Associated with Exit or Disposal Activities, which SEC rules require filing within four business days of the board's commitment to the plan. The SEC's own interpretive guidance allows a company to delay that filing until affected employees have been informed, rather than treat four days as a hard deadline to use in full. That is why the sequence below ends with the filing rather than starting with it.
More employees trust their direct manager most for company updates than trust senior leadership, 52% against 10%, and 86% rely on a manager specifically to translate what a change means for their own role. On Thorncliff's worked rollout, one question the talking points never anticipated, whether the cut reached the two unaffected offices too, got six different improvised answers from six managers in a single day. A same-day fix brought the next five instances to five of five.
Audience SequenceThorncliff Robotics: 242 employees, Austin, Berlin, Bangalore, Sydney. Checked against each audience's own next working hour, not against the external release alone.
| Audience or Asset | Governing Deadline | Gap | Status |
|---|---|---|---|
| Sydney regional lead, live walkthrough | Sydney session, 8:30 AM AEDT | 3.5h | On track |
| Sydney other managers, passive send | Sydney session, 8:30 AM AEDT | 2.5h | Tight |
| Sydney regional session (live) | External release | 16.0h | On track |
| Austin and Berlin 1:1s begin | External release | 1.0h | On track |
| Company-wide all-hands | External release | 0.5h | On track |
| External release (press + Form 8-K) | Reference point | 0.0h | Reference |
| Talking Points v2, all managers | Next local team meeting, every office | 5.0h | On track |
The tightest row is not the all-hands. It is the Sydney regional lead's own briefing material, which needs its own deadline computed separately from the broadcast it supports.
Talking Pointsv1 distributed globally before any live session. Four anticipated questions, one gap found the same day.
| Anticipated in v1 | Result |
|---|---|
| Severance calculation | Consistent from v1 |
| Another round of cuts? | Consistent from v1 |
| Is my own role affected? | Consistent from v1 |
| Does this reach Bangalore or Sydney? | Not anticipated |
What v2 added, same day: "This plan affects only Austin- and Berlin-based roles tied to the product line. It does not change staffing plans in Bangalore or Sydney."
| Bangalore/Sydney question | Instances | Matched |
|---|---|---|
| Before v2 | 6 | 0 (0%) |
| After v2 | 5 | 5 (100%) |
Nobody drafting v1 asked what an unaffected office would want to know. The gap was invisible until two managers in different offices logged the same question on the same day.
Question LogThirty manager-fielded questions across four topics, tagged against whether an approved line existed at the time.
| Question | Instances | Matched | Rate |
|---|---|---|---|
| Reaches Bangalore/Sydney? (before v2) | 6 | 0 | 0% |
| Reaches Bangalore/Sydney? (after v2) | 5 | 5 | 100% |
| Severance calculation | 8 | 8 | 100% |
| Another round of cuts? | 5 | 4 | 80% |
| Is my own role affected? | 6 | 6 | 100% |
| All four questions | 30 | 23 | 76.7% |
The one deviation on "another round of cuts" got a direct note to that manager, not a content revision. The line was already correct; only the delivery was not.
Walks through why a single Austin-clock schedule looked safe in every send-gap number and still failed Sydney, and what a fix on Sydney's own clock actually looks like.
Every office and every distributed asset checked against its own next real working hour, not against headquarters' sequence, with the tightest gap flagged rather than assumed safe.
The one-way broadcast, kept deliberately general and routed to a named next step, since a document nobody can ask a follow-up question to should not try to anticipate every one. The press release and announcement kit is the outward-facing version of the same facts.
Versioned from v1, with every later addition dated and explained, so a manager reading v2 knows what prompted the change instead of assuming v1 was careless.
The same ground as the talking points, written to be read cold, corrected the same day the Question Log finds a gap in either document.
Every manager-fielded question tagged against whether an approved line existed yet, separating a content gap that needs a new line from a manager who skipped a line that already existed. The analyst relations pack routes a similar split, correction request against new evidence, for a different kind of published error.
Hand River the decision documents and whatever external release date or filing deadline already exists, plus the list of offices with employees who need to hear it.
Not just a send gap against headquarters' schedule. An arrival gap per office, and a separate check for any asset, like a regional lead's briefing materials, with its own earlier deadline. The same clock-first check anchors the crisis communication plan, for the version of this problem you do not get to schedule.
The broadcast stays general on purpose. The talking points anticipate what an unaffected office would ask, not only what an affected one would.
A question with no approved line gets one added to both documents. A manager who skipped an existing line gets a direct note instead of a rewritten one.
Yes, with no signup and no card, and the download is the complete pack rather than a preview of it. Edit with AI installs it as a private workspace and sequences your own decision and timeline instead of this worked example. More packs sit in the template library.
Four Word documents and two CSV sheets in one zip. The documents open in Word, Pages, and Google Docs, and the sheets open as real spreadsheets in Excel, Numbers, or Google Sheets, so the sequence and the log are usable data rather than a table flattened onto a page.
The usual template gives a subject line, a bullet list of what's changing, and a timeline table with rows like "manager training" left blank for a date. None of them compute whether that date, converted to a real office's local time, actually falls before the office's next working hour or six hours after it becomes public news elsewhere.
The send gap is the scheduled time before the external release, identical everywhere a single broadcast reaches. The arrival gap is the release time minus that audience's next real working hour after the send, and it is the number that answers whether an office actually hears it from you first.
The FAQ is written to be read cold, with no manager present to add context, so every answer has to stand alone. The talking points are written for a manager to say out loud, versioned from v1, with exact lines marked for anything that has to be word for word, like a severance formula.
It installs this pack as a private workspace and asks for your decision and its external timeline. It builds your own audience sequence from your real offices, drafts the announcement and talking points around what an unaffected group would ask, and logs every question that comes back afterward.
Written by
Co-founder & CEO, River
Chandler has spent the last six years working in AI, building tools for millions of users and Fortune 100 companies. In 2025, he co-founded River to build a new kind of operating system for AI agents.
Send the decision and its external timeline. River checks every office's arrival gap, not just the send gap, and logs what managers actually get asked afterward.
Edit with AI